Fig. 1: Promoter-operator construct and model relating gene expression variability to transcription factor kinetics.
From: Anti-correlation of LacI association and dissociation rates observed in living cells
a The construct consists of a LacI binding site (“Operator”) immediately downstream of the apFAB120 promoter, driving expression of the mVenusNB fluorescent protein. The system is either in a LacI-unbound state (top) in which mRNA is produced at rate r, or a LacI-bound state in which transcription is halted (bottom), (created in BioRender https://BioRender.com/q45e576). mRNA is degraded at rate γ, and ka and kd represent the LacI association and dissociation rates, respectively. Expressions for mean expression μ and Fano factor F as a function of model parameters are shown below the schematic14. b Stochastic simulations of mRNA production for “slow” (top, blue; ka = 3.0 min−1, kd = 0.18 min−1) and “fast” (bottom, orange; ka = 30 min−1, kd = 1.8 min−1) operators. r = 25 min−1 and γ = 0.8 min−1 for both operators. Both operators have the same steady-state LacI binding probability and hence the same mean mRNA expression. However, the “slow” operator exhibits greater variability, reflected in its larger Fano factor and longer-tailed steady-state mRNA copy number distribution. For easier comparison, the “fast” mRNA distribution is plotted (orange dashed line) on top of the “slow” mRNA distribution (blue bars) and vice versa. The mRNA distributions are plotted on a log scale in insets. c Subset of operator mutants assayed in this paper. Operators used the Osym (pink), O1 (green), or O2 (purple) operators as starting points. The wild-type operators are shown in bold and mutations are highlighted in black. Gene expression levels as measured by mRNA FISH (“mRNA”) and fluorescence microscopy (“mVenusNB”) are shown for Osym, O1, and O2. d For each operator, the mean mRNA copy number is plotted against the mean mVenusNB fluorescence, showing (as expected) a strong correlation. Osym, O1, and O2 wild-type operators are plotted with larger circles; error bars represent sem from bootstrapping. Each data point corresponds to at least 862 individual cells assayed across at least two biological replicates. Source data are provided as a Source Data file.
